Skip to content

fix(studio): Monitoring metric deltas compare like with like - #797

Merged
LinuxDevil merged 1 commit into
mainfrom
redesign/r1-monitoring-deltas
Oct 29, 2026
Merged

LinuxDevil merged 1 commit into
mainfrom
redesign/r1-monitoring-deltas

Conversation

@LinuxDevil

Copy link
Copy Markdown
Owner

Monitoring Runs/Cost deltas were skewed low because the current side included the in-progress bucket; compare equal counts of complete buckets, and show the loading state on window switch.

Ticket: R1 (Agent Forge Studio redesign, audit/forge-redesign/PLAN.md)

Changes

  • fix(studio): Monitoring metric deltas compare like with like

Verification: focused vitest, tsc --noEmit, eslint, affected Playwright specs, light + dark screenshots reviewed; full suite and fallow run in CI.

🤖 Generated with Claude Code

The current side of each card's delta included the bucket still in
progress while the previous side was all complete buckets, so Runs and
Cost read low (about -46% on 1h at 10:05). splitWindow now takes the
chart window and returns the whole window (values, sparklines, charts)
plus a comparison of its complete buckets against as many complete
buckets just before `from`; null (no delta) while there is none.

A window or agent switch now shows the cards and charts loading
(useTaggedInsights) instead of the old rows drawn on the new buckets.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@LinuxDevil
LinuxDevil merged commit 631dc0e into main Oct 29, 2026
2 checks passed
@LinuxDevil
LinuxDevil deleted the redesign/r1-monitoring-deltas branch October 29, 2026 13:38
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant